I’ve been thinking about Apple’s developer conference last week and the announcement of “Apple Intelligence” that will be Apple’s branded version of AI. Mostly, AI is about doing intelligent things once you have oodles of data. The big data set that Apple is parsing is data you produce with your phone usage, your photos, your travel, etc. And it’s mostly processing this data locally on your device, not sending your data to a giant supercomputer somewhere else to be queried. This explains partly why Apple started producing massively overpowered devices a few years ago.

I’m excited about these technological developments. Apple demonstrated how the phone could update calendar invites for multiple people based on your current whereabouts. Or how it could intelligently organize your photos into a themed slideshow.

Most of the immediate controversy, flamed by Elon Musk, was about Apple’s contract with ChatGPT for when you specifically request offsite information that ChatGPT may possess.

But, actually, I think the more interesting critique of Apple Intelligence is that there is a flip side to its promise of privacy and security. Because it’s workings in intimate data, and doing so on an intimate device, it will be harder to launch public criticism or assessments of Apple Intelligence. It will be difficult to discern what is a consequence of our own data collection, and what is a consequence of Apple’s programming over it.